Urban-scale ML Engineer for Predictive Weather Systems

Met Office; Exeter; Hybrid; GBP 29,000 - 43,000; Full time

Benefits offered by this job

Hybrid working

Civil Service pension

RRA market supplement

Annual leave

Job summary

The Met Office seeks an Urban-scale Prediction Foundation Scientific Software Engineer to advance 100m-scale urban prediction methods for hazards like thunderstorms, heat and fog. You will build, optimise and deploy ML models within HPC and cloud environments, working from our University of Reading hub with academic partners.

You will contribute to research and software delivery, refactor ML workflows, and help ensure software standards are followed across teams.

Qualifications

• Demonstrable interest or experience in ML/AI methods to solve complex problems.
• Strong written and verbal communication skills.
• Attention to quality assurance, standards and regulations in work.
• Ability to collaborate across teams and partners.

Responsibilities

• Contribute to ML workflows, refactor, modularize and optimize pipelines; assess new algorithms; apply software best practices; optimize performance on HPC/GPUs and cloud.
• Work with Met Office teams to ensure consistency of approach and software standards.
• Collaborate with UK universities and UM partners to advance sub-km/urban prediction initiatives including Reading scientists.
• Document and present work in internal and external venues and provide technical advice to colleagues and stakeholders.

Skills

ML/AI basics

Communication

QA practices

Team collaboration

Job description

The Met Office seeks an Urban-scale Prediction Foundation Scientific Software Engineer to advance 100m-scale urban prediction methods for hazards like thunderstorms, heat and fog. You will build, optimise and deploy ML models within HPC and cloud environments, working from our University of Reading hub with academic partners.

You will contribute to research and software delivery, refactor ML workflows, and help ensure software standards are followed across teams.